Recommendations for Improving the Existing Warning System for an Impending Eruption of the Nevado del Ruiz Volcano, Colombia, South America: An Advance Report
Dennis S. Mileti (Team Leader), Colorado State University, Fort Collins
Patricia A. Bolton, Battelle Memorial Institute, Seattle, Washington
Gabriel Fernandez, University of Illinois, Urbana
Randall G. Updike, Alaska Division of Geological and Geophysical Surveys, Eagle River, Alaska
Committee on Natural Disasters
Commission on Engineering and Technical Systems
National Research Council
NATIONAL ACADEMY PRESS
Washington, D.C. 1986
RECOMMENDATIONS FOR IMPROVING THE EXISTING WARNING SYSTEM FOR AN IMPENDING ERUPTION OF THE NEVADO DEL RUIZ VOLCANO, COLOMBIA, SOUTH AMERICA: AN ADVANCE REPORT
Shortly after the 13 November 1985 eruption of the Nevado del Ruiz volcano in Colombia, South America, the Committee on Natural Disasters of the Commission on Engineering and Technical Systems, National Research Council, organized and dispatched a four-person study team to gather data about the eruption, its consequences, and the warning and recovery capabilities available to deal with future eruptions. The team is currently in the process of preparing a full report of its findings. However, the team has also decided to prepare this advance report on the warning system in Colombia to make their findings available without delay to those officials who can take immediate action to correct the deficiencies of the current warning system.
Geological and seismological data on the activities of the volcano gathered at the Manizales Observatory indicate that the volcano has an extremely high probability of another major eruption soon. The only question is how soon and how “major.” The impending eruption of the volcano has the potential to be a major disaster that could kill even more people than the 22,000 to 24,000 people killed in the 13 November 1985 eruption. Using population data and information from township maps and risk maps, an estimated 50,000 to 80,000 additional people are at risk of losing their lives when the volcano again erupts.
An effective warning-evacuation system could substantially reduce the risk of a disaster occurring comparable to that of November 1985.
However, the existing system is not considered effective because of two significant flaws: (1) the system may not be fast enough to pass a warning message to those who need to evacuate, and (2) the type of warning issued may not be effective enough to convince those who should evacuate to do so. Overcoming these flaws requires that immediate and concentrated attention be given to transmitting information quickly and effectively from scientists monitoring the volcano to appropriate government officials in order to promptly warn the people who are at risk from an impending eruption.
THE VOLCANO AND THE DISASTER EVENT
The Nevado del Ruiz volcano, with a maximum elevation of 5,400 m, is one of the high peaks of the central cordillera of Colombia. It is located about 150 km west-northwest of Bogotá and about 30 km southeast of the city of Manizales, which has a population of 350,000. The volcano has been quiet for nearly the past 400 years. The last major eruption, about which few details are available, occurred in 1595. In 1845 the volcano was the source of an avalanche that descended the eastern slope of the mountain and caused mudflows to reach the Magdalena River more than 60 km away, killing over 1,000 people.
At about 9:00 p.m. on 13 November 1985, two sudden blasts followed by a 25-minute eruption of red-hot pumice blocks from the crater of the Nevado del Ruiz volcano melted part of the ice cap that crowns the volcano. The resulting mudflow--mixture of water, ice, pumice, and soil--sped down the mountainside, at speeds reaching approximately 50 km/hr, via the Azufrado River channel and the Lagunilla River. The mudflow had such force that it caused the collapse of a natural dam on the Lagunilla River and swept away the town of Armero, located about 45 km east of the volcano. Between 22,000 and 24,000 people at Armero perished.
The mudflow also caused the Gaulí River to overflow, which carried away houses and a bridge on the main road to Bogotá, located about 150
km east-southeast of the volcano. Another mudslide descended along the west bank of the volcano, reaching the town of Chinchina. It destroyed about 400 houses and caused an estimated 1,000 deaths.
THE STUDY TEAM’S TRIP
The National Research Council, through its Committee on Natural Disasters, Commission on Engineering and Technical Systems, dispatched a four-person team in February 1986 to Colombia to study the eruption and its aftermath. The team stayed in Columbia from 9 through 14 February 1986. The team’s charge was:
To provide a reasonably accurate and conveniently available account of the event for historical purposes.
To study the characteristics of the debris flow (i.e., its velocity, direction, areal coverage, and composition and particle distribution with distance from the origin).
To identify and recommend areas where additional research could contribute to the improvement of preparedness, warning, evacuation, rescue, recovery, and rehabilitation capabilities.
The team’s field trip was extremely productive, especially in the compilation of data on emergency preparedness, response planning, and geological and geotechnical engineering. The following topics will be included in the team’s final comprehensive report:
The disaster and the events leading up to it
Preeruption public education and emergency preparedness
The warning system for, and public response to, the 13 November 1985 eruption
The immediate disaster response and search and rescue
Relief and rehabilitation, including the provision of temporary housing and aid
Reconstruction of permanent housing for survivors
Short- and long-term consequences of the event
However, at least six months will be required for preparation, review, and distribution of this report.
WHY AN ADVANCE REPORT?
The 13 November 1985 event was a disaster of immense proportions even though the size of eruption was quite small. Geologists estimate that the volcano has an extremely high probability of a major eruption in the near future. Using population data and information from township maps and risk maps, along with geological and seismological data available to the team members during their stay, an estimated 50,000 to 80,000 additional people will be at risk of losing their lives.
The preparations for the impending eruption of the volcano are matters of utmost urgency. The record of the events of 13 November 1985 confirms that the disaster at Armero resulted not from a failure of scientists to warn of the impending eruption but from breakdowns in attempts to communicate that information to those who could have evacuated. Repeated orders for the evacuation of Armero were constrained from being fully received in the town, were not fully implemented, and never reached many members of the public.
Despite the sincere and elaborate efforts of many dedicated Colombians, the warning-evacuation system designed to save the lives of the 50,000 to 80,000 people currently at risk from the volcano’s next eruption is flawed. Left uncorrected, these flaws could result in human deaths of staggering proportions when the volcano erupts. Such a disaster would be extremely unfortunate given that these flaws in the warning system could be corrected through the application of existing knowledge about the warning-evacuation process. For this reason, the study team chose to prepare an advance report in order to make its findings available without delay to those officials who can take action to correct the current deficiencies.
THE EXISTING WARNING SYSTEM
The warning-evacuation system designed for the next eruption is flawed in two ways. First, the warning system may not be fast enough to get a warning message to those who need to evacuate. Second, the type of warning issued may not be effective enough to convince those who should evacuate to do so.
The existing warning system in Colombia is designed to pass warnings from scientists to the public through many governmental and bureaucratic levels. The system was activated on 4 January 1986 when a minor eruption occurred marked by ash emitted from the summit. Fortunately, no deaths occurred since this eruption was not accompanied by the emission of pyroclastic flows, which triggered the disaster of 13 November 1985.
In the January event it took three hours for the warning message to reach the public. However, some endangered communities could be affected within 45 minutes of an eruption, and most towns at risk would be affected within 2-1/2 hours. This delay could result in obvious problems of protecting endangered communities in the event of a future major eruption.
Existing plans call for initiating public evacuation by evacuation warnings, including sirens. Extensive evidence from past evacuation research demonstrates that sirens, even when preceded by elaborate preevacuation public education, are not enough to get people to leave their homes. What to say, how to say it, and what not to say is well understood by the research community for emergency planning and is of dramatic importance in increasing the probability of effective evacuations. The evacuation warning issued because of the 4 January 1986 event resulted in little actual evacuation.
If scientists can predict an impending volcanic eruption of Nevado del Ruiz within a specific time frame--e.g., several days ahead of time--there could be enough time (even without emergency plans) to inform the people at risk and evacuate them to safety. Since this is still not possible and the best scientists can do is to indicate the extremely high probability of the impending event, the implementation of emergency warnings and evacuations will be critical in saving lives.
Volcanic eruptions represent the type of risk that, with modern monitoring and careful preparedness planning, can be significantly mitigated. Moreover, the cost of such predisaster preparedness is low compared with the benefit of saving lives.
Colombia’s warning system could benefit from several measures:
Additional instruments should be installed for monitoring potential mudflows.
There is an urgent need to install additional instruments at critical locations to monitor the advance of mudflows in the event of a future volcanic eruption. The telemetered seismic and ground deformation monitoring networks now deployed around the volcano give scientists at the Manizales Observatory immediate notification that an eruption is under way. However, except in a few river valleys, the networks cannot confirm that an eruption has actually generated mudflow activity. In addition, the networks may not be able to detect mudflows that have been triggered by noneruptive mechanisms, such as a landslide or other event. A mudflow detection network, deployed in all of the principal rivers that originate in the snowfields of the volcano, is needed to ensure that no potentially damaging mudflow occurs without public evacuation warnings.
The time it takes to inform the public at risk should be reduced to much less than three hours.
Mudflow experts have calculated that mudflows could travel from the volcano to Ambalema or Mariquita, which are about 50 km east and north-east of the volcano, respectively, in about an hour. The time to get a first alert to the public at risk must therefore be reduced to well
below one hour. This could be achieved in many different ways. All possibilities involve both technological and preparedness planning components.
In the technological area, a redundant and direct communication system should be established between the scientific monitoring stations on the volcano’s western slope and the offices of the President and Civil Defense headquarters in Bogotá. At present, the warning system consists of a direct telephone line between these offices with a battery-operated telephone line as the backup. Consideration should be given to increasing the redundancy of this link by installing a microwave telephone system at the scientific monitoring station. The microwave telephone system is an existing system that provides a direct communication capability between the President and provincial governors in the case of a state emergency.
The option of establishing a reliable means of communication between these offices using satellite communication technology should also be examined. Colombia’s membership in Intelsat (the International Telecommunications Satellite Organization) offers it a good opportunity for creating such a system.
If the satellite option is selected, the earth stations and terrestrial rebroadcast stations will have to be either automated or staffed by reliable, well-trained personnel. The former may be preferable, with the stations on keep-alive standby until activated by an emergency signal from the Presidential transmission. A trained cadre of reliable technicians would be required to periodically check and maintain the earth stations.
The President and the head of Civil Defense should have a direct link through any of the communication systems described above to battery-operated radios, sirens, and loudspeakers throughout all of the communities at risk. Officials backing up the President and the head of Civil Defense, who have authority to act in their stead, should also be equipped with the same communication system. The public in the villages could be provided with inexpensive battery-operated radio receivers: the cheapest type would be single-frequency tuned to the emergency
broadcasting frequency. Other options also exist that would ensure that members of the endangered public would hear warning messages.
Preparations should be made to give on-the-spot public information based on what is known about verbal evacuation warnings.
In the planning and preparedness area, planners should catalog the range of potential scenarios that scientists could send to the President and the head of Civil Defense. In concert with provincial governors, mayors from local communities at risk, and other influential local officials, planners should decide when public evacuation would be recommended on the basis of different scientific communications. These “response” scenarios could then be exercised during the planning stage without including the public so that it is clear what scientific events could precipitate the execution of immediate evacuation warnings. The participation of local government is important in any planning effort so that local officials can be bypassed to save time in communicating warnings to the public when the system is activated.
If a response scenario calls for sirens to be activated to alert the at-risk population, then activation of the sirens should be immediately followed by multiple and repeated verbal warnings from a known public official (perhaps the President) directly to the people who should evacuate. These should be followed by further soundings of sirens and verbal warnings through loudspeakers at local levels. Local government and Civil Defense officials, and perhaps even Red Cross personnel, should plan to ensure public compliance with the evacuation warnings. Plans should be developed that would facilitate these officials’ ensuring that the public is correctly interpreting and acting upon the verbal warning messages received.
The content of the verbal warnings given to people at risk should be decided ahead of time. This content should be shaped by what has been learned about emergency evacuation warnings through more than 25 years of research and practice in nations like the United States and Japan.
Preparation of verbal warnings should adhere to the following principles, each of which has been shown to increase the odds of people actually evacuating:
Source--it should be a credible source who states that the warning is endorsed by the government and scientists.
Consistency--the message should not contradict itself (e.g., “the volcano erupted but don’t worry”).
Accuracy--the message should present the best scientific evidence about the possible risk and not forget to tell people why they should evacuate.
Clarity--the message should be drafted in words familiar to those who will need to evacuate and not in words that only officials and scientists are comfortable with.
Certainty--the message should display verbal confidence in the voice and tone of the speaker.
Guidance--people must be told exactly what to do, where to go, and how long they have to follow these instructions regardless of the extent of preemergency public education efforts.
Level of information--the message must inform people about the details of what is going on (e.g., concerning the eruption, mudflow, height of mud, what will happen to their town, which towns are being evacuated).
Frequency--the message must be repeated frequently.
Location--the message must clearly state who is at risk and who must evacuate in such a way that those at risk will perceive clearly that the message is directed to them.
A half dozen or so example messages (conforming to the above guidelines) should be developed. More than one are needed because alternative eruption scenarios (a preevent false alarm, another minor eruption, etc.) would require different messages regarding evacuation.
Existing electrical supplies should not be relied upon.
Electrical power is essential to operate the existing telephone lines, the earth stations in a satellite communication system, and the receivers. A reliable power source could be achieved in part by providing on-site diesel generators with their own on-site fuel supplies sufficient to run these communication systems for whatever warning period is considered appropriate plus a four- to six-times reserve. Power for the public’s receivers would almost certainly have to come from batteries installed in the receivers themselves or from solar power. A proper testing schedule should also be established and exercised to periodically check the reliability of these power supply systems to ensure their operability in the event of a disaster.
A number of issues must be addressed in order to implement these recommendations. These issues may be grouped into three major categories: (1) technical hardware for mudflow monitoring, communications from scientists to officials to the public, and backup electrical supply; (2) emergency preparedness and exercises to provide for timely public alert, the delivery of sound verbal warning messages to the public, and local official monitoring of public actions in response to warning messages; and (3) the financial and personnel resources needed to achieve these two objectives.
Technical capabilities for installation of proper instruments for mudflow monitoring can best be obtained from the present international team that is monitoring the activities of the volcano, including geologists from both Colombia and the U.S. Geological Survey. The study team also recommends that Colombia officials consult immediately with a group of experts to resolve issues related to the communication systems best suited for the local situation and the emergency planning needed to upgrade the existing warning system for the impending volcanic eruption. The emergency planning staff of the Long Island Lighting Company has indicated their willingness to assist the group of experts in offering technical and planning assistance to Colombia if called upon to do so.